Re: SCO's infringing files list

[Bruno Haible]

Here's some info about include/linux/ipc.h, also in SCO's list.

* lxr.linux.no shows that since version 1.0.9, it had only small incremental
  changes.

The earliest copy of this file that I've got is from Krishna
Balasubramanian's ipcbeta+.tar.Z file. This was his second or third
beta release of SysV IPC for Linux. The file is here:
http://www.haible.de/bruno/ipcbeta+.tar.Z

* The include/linux/ipc.h from ipcbeta+.tar.Z is the same as the one in
  linux-1.0.9 (http://lxr.linux.no/source/include/linux/ipc.h?v=1.0.9)

I claim that Krishna Balasubramanian wrote this file.

* The ipcbeta+.tar.Z contents shows how he developed this thing:
  He looked at various documentation sources (books, manual pages - remember
  POSIX didn't specify IPC at that time -).
  He collected some examples like the "dining philosophers" that were floating
  around on the net.
  We ran some test programs on other Unices (SunOS 4, possibly also HP-UX).
  He wrote 40 KB of documentation, explaining each and every system call.
  ... and someone who puts so much work in testing and documentation should
  steal the header file??!

* The value of IPC_PRIVATE is different in Linux. SysV systems define it as
  (key_t)0, Linux defines it as ((key_t) 0), which extra parentheses.

* The members of 'struct ipc_perm' are in different order on Linux.
  SysV systems have them in the order
     uid, gid, cuid, cgid, mode, seq, key.
  Linux has them in the order
     key, uid, gid, cuid, cgid, mode, seq.

* The values for IPC_CREAT, IPC_EXCL, IPC_NOWAIT are written as octal numbers,
  which is quite natural, since 9 bits having the same rwxrwxrwx semantics as
  file permissions can be ORed into it. SysV systems write these constants
  with 7 octal digits. Linux ipc.h writes them with 8 octal digits.

* The values of IPC_RMID, IPC_SET, IPC_STAT are different: on Solaris
  10, 11, 12; on Linux 0, 1, 2.

* 'struct ipc_kludge' and the corresponding #defines for SEMOP, SEMGET etc.
  don't exist in SysV systems. They arose only because we wanted to minimize
  the number of system calls.

I hope that's enough evidence that Krishna didn't copy the file's contents
from anywhere.

Bruno

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/